Desempenho do modelo APSIM em crescimento e produtividade de milho em Alagoas

Abstract

The crop modelling is a support tool in decision making and management of agricultural cropping systems under specifics conditions. The Agricultural Systems SIMulator (APSIM) is a model of agricultural systems, structured in biological and environmental modules, and allows users to create simulations with complex management pratices and under different environmental conditions. The maize (Zea mays L.) was chosen to exemplify the simulations performed in this study due to its wide use in various worldwide socioeconomic sectors and also stands out among the crops most produced in Brazil. In the State of Alagoas, Northeastern Brazil, the maize is relevant mainly for family farms, and has productivity below the national average due to low technology used in cropping systems. However, the use of crop models, associated with higher levels of technology and appropriate management, can contribute to the optimization of productivity in region. Thus, the aim of this study, was evaluate the APSIM-Maize model to simulate growth and productions variables of maize in the State of Alagoas, Brazil. The regions studied were Arapiraca-AL and Rio Largo-AL, in which the model was evaluated with edaphoclimatic data of regions. Initially were inserted phenological variables relating to maize variety used experimentally (AL Bandeirante), soil water parameters and physicochemical parameters for the model to identify and reproduce the soil water dynamics and growth process of the crop. The APSIM-Maize simulated the soil water content with average difference between observed and simulated data (RMSE) ranging between 0.021 and 0.053 mm mm-1, wherein, the estimates for the Arapiraca-AL were better adjusted with observed data. For simulations in Rio Largo-AL, calculations of soil water dynamics were influenced by a physical characteristic of the soil (dense layer of soil at a depth of 30 cm) that was identified, but not well represented by the model. This edaphic factor was minimized by adjustments in the drainage parameter. The leaf area index (LAI) was estimated with RMSE of 0.69 cm2 cm-2, and in most cases adequately reproduces the observations until about 60 days after sowing. In some sowing dates, the simulations of the LAI was efficient throughout the crops cycle with low percentage deviations. Biomass production was simulated with RMSE between 2.04 and 3.33 Mg ha-1, and had large deviations during the growth cycle of the culture, however for the final biomass production values the model has satisfactory estimates with percentage errors varying between 3.52 and 35.16%. Therefore, the APSIM-Maize proved to be a reliable model to estimate the components of the water balance in the soil. But showed inconsistencies in some cases to estimate growth and production variables for the conditions studied. The calibration presented here is only valid for edaphoclimatic conditions for which the model has tested, i.e., the model can not be applied to the Alagoas State as a whole, given that the APSIM model is sensitive to environmental changes and requires a specific calibration for each location.